There is enormous variation in medical spending and use of services across geographic areas in the United States. A large portion of this variation has been linked to use of services that have not been studied in randomized trials. This wide variation in services of unknown benefit has led to calls for research using observational data to measure the comparative effectiveness of medical treatments. Estimating treatment effectiveness in non-randomized settings requires careful data collection and analysis because of possible unobserved differences between patient groups. Increasingly, researchers are exploiting geographic variation in use of services to infer their effectiveness in individual patients, under the assumption that patient characteristics are balanced across geographic areas. However, little is understood about settings in which ecological variation can be used to obtain valid estimates of treatment effects in non-randomized studies. Using data from population-based cohorts of elderly patients diagnosed with colorectal and prostate cancer in the Surveillance Epidemiology, and End Results (SEER)-Medicare database, we will develop, evaluate, and apply advanced statistical methods to estimate the effectiveness of cancer interventions. Cancer provides an ideal setting for such an analysis because it is a prevalent and serious health condition, and new and expensive treatments of unknown benefit are continuously being developed and diffused. These goals will be accomplished through the following three specific aims: (1) develop and evaluate statistical methods to estimate treatment effectiveness using geographic residence to mimic randomization to treatment;(2) apply these methods to estimate the comparative effectiveness of interventions for patients diagnosed with colorectal and prostate cancer;and (3) develop and disseminate software to promote the use of appropriate statistical methods in comparative effectiveness studies These results will provide information about the effectiveness of interventions highlighted in a recent Institute of Medicine study on priorities for comparative effectiveness research and also provide a methodology that other researchers can use to learn about the effectiveness of treatments in other diseases. PUBLIC HEALTH RELEVANCE: This project will provide valuable information on interventions in colorectal and prostate cancer highlighted in a recent Institute of Medicine study on priorities for comparative effectiveness. In addition, the methods developed will be applicable to a broad range of interventions for cancer and other diseases, providing a framework for future comparative effectiveness studies and ultimately changes in treatment guidelines.